(a) (3 points) In the K-Means algorithm, the initial placement of centroids has a minimal impact on the quality of the resulting clustering. (b) (3 points) The Expectation Maximization E-M algorithm will always converge to global minimum. (c) (3 points) Suppose two association rules are derived from the same frequent pattern; the two association rules must share the same support and confidence. (d) (3 points) Given two itemsets A and B, the range of Kulczynski(A,B) measure of A and B is (0, 1). (e) (3 points) "c : average(S.profit) >= 50 " is pattern anti-monotone. (f) (3 points) Let p and q be two different distributions. The KL divergence between p and q is the same as the KL divergence between q and p. (g) (3 points) For a perfect clustering, the conditional entropy value is 1. (h) (3 points) After applying a z-score standardization on a group scalars, their values will fall into the range of [-1, 1